{tabs.map((tab) => (
))}
{/* ================================================================= */}
{/* TAB 1 — Overview */}
{/* ================================================================= */}
{activeTab === 'overview' && (
{/* Metric Cards */}
{/* Platform Highlights */}
)}
{/* ================================================================= */}
{/* TAB 2 — Data Pipeline */}
{/* ================================================================= */}
{activeTab === 'pipeline' && (
{[
{ label: 'Products in Catalog', value: EAGLE_DATA.overview.productsInCatalog, suffix: '+', icon:
{/* Mission Cluster Distribution */}
{m.icon}
{m.label}
Mission Cluster Distribution
6 shopper mission segments derived from purchase pattern clustering
{EAGLE_DATA.missionClusters.map((seg) => (
{seg.name}
{seg.count.toLocaleString()}
{seg.pct}% of shoppers
{seg.desc}
Data Foundation
4 tables
Delectable Data Platform tables power every agent conversation: shopper profiles, product catalog, purchase history, and product embeddings.
Personalization Depth
11 dietary signals
Per-shopper propensity scores across 11 dietary dimensions power proactive filtering and alternative suggestions.
Multi-Aspect Search
3 embedding types
Identity, nutrition, and description embeddings enable nuanced product discovery beyond simple keyword matching.
{/* Source Tables */}
{/* Entity Connection Graph */}
)}
{/* ================================================================= */}
{/* TAB 3 — Customer 360 (Delectable AI Shopper Profile) */}
{/* ================================================================= */}
{activeTab === 'customer360' && (
Delectable Data Platform Source Tables
{EAGLE_DATA.pipeline.tables.map((t) => (
))}
{t.icon === 'users' &&
{t.name}
{t.rows}
{t.desc}
Entity Connection Graph
Data flows from shopper profiles through purchase history and product catalog into embeddings, which the agent queries via VECTOR_SEARCH.
Data Volume Summary
{[
{ label: '135K Shoppers', color: 'bg-blue-600' },
{ label: '70K Products', color: 'bg-emerald-600' },
{ label: '48M Transactions', color: 'bg-orange-600' },
{ label: '768-dim Embeddings', color: 'bg-purple-600' },
{ label: '2yr History Window', color: 'bg-pink-600' },
{ label: '11 Dietary Signals', color: 'bg-cyan-600' },
].map((b) => (
{b.label}
))}
Join Key Details
| Source | Target | Join Key | Match Rate | Purpose |
|---|---|---|---|---|
| {conn.source} | {conn.target} |
{conn.joinKey}
|
|
{conn.source === 'Shoppers' && conn.target === 'Purchase History' && 'Link shopper to transaction history'} {conn.source === 'Purchase History' && conn.target === 'Products' && 'Resolve purchased products'} {conn.source === 'Products' && conn.target === 'Embeddings' && 'Multi-aspect vector representation'} {conn.source === 'Embeddings' && conn.target === 'Agent' && 'Semantic search at query time'} {conn.source === 'Shoppers' && conn.target === 'Agent' && 'Real-time profile + pantry lookup'} |
{/* Shopper Header */}
{/* Dietary Propensity Scores */}
{/* Virtual Pantry */}
{/* AI Insight */}
)}
{/* ================================================================= */}
{/* TAB 4 — Agent Analytics */}
{/* ================================================================= */}
{activeTab === 'agentAnalytics' && (
Delectable AI Shopper Profile
Shopper ID
{EAGLE_DATA.shopper.id}
Mission Cluster
{EAGLE_DATA.shopper.missionCluster}
Cluster {EAGLE_DATA.shopper.clusterNumber} of 6
Loyalty Tier
{EAGLE_DATA.shopper.loyaltyTier}
Dietary Dimensions
11 propensity scores
Dietary Propensity Scores
Per-shopper scores (0-1) derived from 2-year purchase history
{EAGLE_DATA.shopper.dietaryScores.map((d) => {
const barColor = d.score >= 0.7 ? '#10b981' : d.score >= 0.4 ? '#f59e0b' : '#64748b';
return (
);
})}
{d.label}
{d.score.toFixed(2)}
{d.score >= 0.7 && HIGH}
{d.score >= 0.4 && d.score < 0.7 && MED}
{d.score < 0.4 && LOW}
Virtual Pantry Snapshot
Estimated stock levels based on purchase frequency and typical consumption rates
{EAGLE_DATA.shopper.pantry.map((item) => {
const stockColor = item.stock >= 0.5 ? '#10b981' : item.stock >= 0.2 ? '#f59e0b' : '#ef4444';
const stockLabel = item.stock >= 0.5 ? 'In Stock' : item.stock >= 0.2 ? 'Running Low' : 'Likely Out';
return (
{item.name}
{stockLabel}
High propensity for organic (0.91), gluten_free (0.78), and high_protein (0.89). The agent will prioritize organic products in search results, flag gluten-containing items with warnings, and proactively suggest protein-rich alternatives. Virtual pantry shows Greek Yogurt and Chicken Breast are likely depleted — these will be auto-suggested for replenishment.
{/* Top-level KPIs */}
{/* Tool Call Distribution */}
{/* Logging Tables Summary */}
)}
{/* ================================================================= */}
{/* TAB 5 — Vector Search */}
{/* ================================================================= */}
{activeTab === 'vectorSearch' && (
{[
{ label: 'eagle_ai_searches', value: EAGLE_DATA.agentAnalytics.searches, sub: 'Total searches logged', color: 'from-blue-500 to-blue-600' },
{ label: 'eagle_ai_selections', value: EAGLE_DATA.agentAnalytics.selections, sub: `${EAGLE_DATA.agentAnalytics.addToCartRate}% add-to-cart rate`, color: 'from-emerald-500 to-emerald-600' },
{ label: 'eagle_ai_conversations', value: EAGLE_DATA.agentAnalytics.conversations, sub: `Avg ${EAGLE_DATA.agentAnalytics.avgTurns} turns / convo`, color: 'from-purple-500 to-purple-600' },
{ label: 'Personalization Rate', value: EAGLE_DATA.agentAnalytics.personalizationRate, suffix: '%', sub: 'Use at least 1 signal', color: 'from-teal-500 to-teal-600' },
].map((m) => (
{/* Conversation Stats */}
{m.label}
{m.sub}
Avg Turns per Conversation
{EAGLE_DATA.agentAnalytics.avgTurns}
User messages per session
Avg Tool Calls per Conversation
{EAGLE_DATA.agentAnalytics.avgToolCalls}
Agent function invocations
Add-to-Cart Rate
{EAGLE_DATA.agentAnalytics.addToCartRate}%
Selections resulting in cart add
Tool Call Distribution
{EAGLE_DATA.agentAnalytics.toolBreakdown.map((t) => (
))}
{t.tool}
{t.pct}%
{t.calls.toLocaleString()}
Platform Logging Tables
{[
{
table: 'eagle_ai_searches',
rows: '48,200',
desc: 'Every product search the agent executes, with query text, embedding type used, result count, and latency.',
fields: ['search_id', 'shopper_id', 'query', 'embedding_type', 'result_count', 'latency_ms'],
color: 'border-blue-500',
},
{
table: 'eagle_ai_selections',
rows: '31,400',
desc: 'Products selected or added to cart from agent suggestions, including position rank and personalization signals.',
fields: ['selection_id', 'search_id', 'product_id', 'position', 'added_to_cart', 'signals_used'],
color: 'border-emerald-500',
},
{
table: 'eagle_ai_conversations',
rows: '12,400',
desc: 'Full conversation metadata: turn count, tool calls, duration, and personalization coverage.',
fields: ['conversation_id', 'shopper_id', 'turns', 'tool_calls', 'duration_sec', 'personalized'],
color: 'border-purple-500',
},
].map((t) => (
))}
{t.table}
{t.rows} rows
{t.desc}
{t.fields.map((f) => (
{f}
))}
{/* Embedding Type Cards */}
{/* Interactive Search Demo */}
)}
{/* Example Queries with Multi-Aspect Scores */}
);
})()}
{/* How It Works */}
)}
{/* ================================================================= */}
{/* TAB 5.5 — Embeddings Projector (3D point cloud) */}
{/* ================================================================= */}
{activeTab === 'embeddingsProjector' && window.EmbeddingsProjector && (
Three complementary embedding types power nuanced product discovery via Delectable VECTOR_SEARCH
{EAGLE_DATA.vectorSearch.embeddingTypes.map((emb) => (
{emb.dim}d vector
Fields: {emb.fields}
{emb.purpose}
Try a Search
setSearchQuery(e.target.value)}
onKeyPress={(e) => e.key === 'Enter' && handleSemanticSearch(searchQuery)}
placeholder='Try: "marinated chicken breast", "high-protein snack", "pasta dinner"'
className="flex-1 bg-white/10 border border-white/20 rounded-lg px-4 py-3 text-white placeholder-gray-500 focus:outline-none focus:ring-2 focus:ring-teal-500"
/>
{searchLoading && (
Querying Delectable AI agent endpoint...
)}
{searchResults.length > 0 && !searchLoading && (
Semantic Search Results:
{searchSource && (
{searchSource === 'live' ? 'Live API' : 'Demo Fallback'}
)}
{searchResults.map((r, i) => (
{r}
))}
Multi-Aspect Search Results
{EAGLE_DATA.vectorSearch.exampleQueries.map((eq, i) => (
))}
{(() => {
const eq = EAGLE_DATA.vectorSearch.exampleQueries[activeVectorQuery];
return (
Primary embedding:
{eq.primary}
| Product | Identity | Nutrition | Description |
|---|---|---|---|
|
{ri + 1}
{r.name}
|
{['identity', 'nutrition', 'description'].map((dim) => {
const val = r[dim];
const isPrimary = dim.charAt(0).toUpperCase() + dim.slice(1) === eq.primary;
return (
|
);
})}
How It Works
{[
{ step: '1', title: 'User Query', desc: 'Natural language search input from the shopper', color: 'bg-gray-600' },
{ step: '2', title: 'text-embedding-005', desc: 'Query converted to 768-dim vector via Delectable AI Platform', color: 'bg-blue-600' },
{ step: '3', title: 'VECTOR_SEARCH', desc: 'Cosine similarity across 3 embedding types', color: 'bg-purple-600' },
{ step: '4', title: 'Dietary Re-rank', desc: 'Results re-ranked by shopper dietary propensities', color: 'bg-emerald-600' },
{ step: '5', title: 'Results', desc: 'Personalized, scored product recommendations', color: 'bg-teal-600' },
].map((s, i) => (
{s.step}
{s.title}
{s.desc}
{i < 4 && (
{/* ── Coverage Impact Hero ──────────────────────────────────────── */}
{/* Pagination controls */}
{unenrichedTotalPages > 1 && (
)}
{/* ================================================================= */}
{/* TAB 7 — Agent Context */}
{/* ================================================================= */}
{activeTab === 'agentContext' && (
Catalog Enrichment Impact
How Delectable AI transformed Acme Grocer's raw product catalog into an intelligence-rich data asset
Catalog Enriched
{EAGLE_DATA.catalogEnrichment.coverageImpact.enrichedCount.toLocaleString()} of 70,000 products
Avg New Data Points
Per product across all enrichment layers
Products Untouched
{EAGLE_DATA.catalogEnrichment.coverageImpact.untouchedPct}% — mostly discontinued or pending
{EAGLE_DATA.catalogEnrichment.coverageImpact.untouchedReasons.map((r) => (
))}
{/* Before / After Comparison */}
{r.count.toLocaleString()}
{r.reason}
Before Enrichment
{EAGLE_DATA.catalogEnrichment.coverageImpact.beforeAfter.before.avgFields}
avg fields per product
{(EAGLE_DATA.catalogEnrichment.coverageImpact.beforeAfter.before.totalDataPoints / 1000).toFixed(0)}K
total data points
After Enrichment
{EAGLE_DATA.catalogEnrichment.coverageImpact.beforeAfter.after.avgFields}
avg fields per product
{(EAGLE_DATA.catalogEnrichment.coverageImpact.beforeAfter.after.totalDataPoints / 1000000).toFixed(2)}M
total data points
Enrichment Layers
Six progressive layers of intelligence applied to the catalog — each layer builds on the previous
{EAGLE_DATA.catalogEnrichment.enrichmentLayers.map((layer) => (
))}
{layer.layer}
{layer.name}
{layer.desc}
{layer.count.toLocaleString()}
{((layer.count / 70000) * 100).toFixed(0)}%
Catalog Composition & Enrichment Tiers
Acme Grocer's catalog spans food, health & wellness, and non-food — each tier receives tailored enrichments
{EAGLE_DATA.catalogEnrichment.catalogBreakdown.map((tier) => (
{tier.tier}
{tier.pct}%{tier.count.toLocaleString()} products
{tier.enrichments.map((e) => (
{e}
))}
{tier.examples.map((ex) => (
))}
{ex.name}
{ex.dataPoints} pts
{ex.note}
How Enrichments Drive Value
Each enrichment layer translates directly into measurable business outcomes across personalization, advertising, and operational efficiency
{EAGLE_DATA.catalogEnrichment.valueDrivers.map((driver) => (
))}
{driver.name}
{driver.desc}
{driver.metric}
Enrichment Explorer — Before & After
Browse the top enriched products by enrichment gain, or search for a specific product to see a side-by-side comparison of raw catalog data vs. enriched intelligence.
{/* Search + Sort controls */}
setEnrichmentSearch(e.target.value)}
placeholder="Search by product name, brand, category, or GTIN..."
className="flex-1 bg-white/10 border border-white/20 rounded-lg px-4 py-2.5 text-white text-sm placeholder-gray-500 focus:outline-none focus:ring-2 focus:ring-emerald-500"
/>
{enrichmentSearch && (
)}
{/* Results count */}
Showing {Math.min(enrichmentPage * ENRICHMENT_PAGE_SIZE + 1, filteredEnriched.length)}–{Math.min((enrichmentPage + 1) * ENRICHMENT_PAGE_SIZE, filteredEnriched.length)} of {filteredEnriched.length} products
{enrichmentSearch && matching “{enrichmentSearch}”}
Page {enrichmentPage + 1} of {enrichmentTotalPages}
{paginatedEnriched.length === 0 ? (
{/* Pagination */}
{enrichmentTotalPages > 1 && (
No products match your search
) : paginatedEnriched.map((product) => {
const isExpanded = expandedProduct === product.gtin;
const beforeKeys = Object.keys(product.before);
const afterKeys = Object.keys(product.after);
const newKeys = afterKeys.filter(k => !beforeKeys.includes(k));
const tierColors = {
food: { bg: 'bg-emerald-500/20', text: 'text-emerald-400', label: 'Food & Grocery' },
health: { bg: 'bg-purple-500/20', text: 'text-purple-400', label: 'Health & Wellness' },
nonFood: { bg: 'bg-amber-500/20', text: 'text-amber-400', label: 'Non-Food' },
};
const tc = tierColors[product.tier] || tierColors.food;
return (
{/* Card header */}
{/* Expanded: side-by-side before / after */}
{isExpanded && (
)}
);
})}
{/* BEFORE column */}
{/* AFTER column */}
Before Enrichment
{beforeKeys.length} fields
{beforeKeys.map(key => (
{key}
{typeof product.before[key] === 'object'
? JSON.stringify(product.before[key])
: String(product.before[key])}
))}
After Enrichment
{afterKeys.length} fields
+{newKeys.length} new
{afterKeys.map(key => {
const isNew = !beforeKeys.includes(key);
const val = product.after[key];
return (
{key}
{typeof val === 'object'
? (Array.isArray(val) ? val.join(', ') : JSON.stringify(val))
: String(val)}
{isNew && new}
);
})}
{Array.from({ length: enrichmentTotalPages }, (_, i) => (
))}
)}
Unenriched Products Explorer
{EAGLE_DATA.catalogEnrichment.coverageImpact.untouchedCount.toLocaleString()} products were not enriched. Browse by reason, category, brand, or keyword to understand why and identify remediation paths.
{/* Reason breakdown cards */}
{EAGLE_DATA.catalogEnrichment.unenrichedProducts.reasons.map((r) => (
))}
{/* Active reason description */}
{unenrichedFilter !== 'all' && (() => {
const active = EAGLE_DATA.catalogEnrichment.unenrichedProducts.reasons.find(r => r.code === unenrichedFilter);
return active ? (
{active.label}
{active.desc}
setUnenrichedSearch(e.target.value)}
placeholder="Search by product name, brand, category, or GTIN..."
className="flex-1 bg-white/10 border border-white/20 rounded-lg px-4 py-2.5 text-white text-sm placeholder-gray-500 focus:outline-none focus:ring-2 focus:ring-amber-500"
/>
{(unenrichedFilter !== 'all' || unenrichedSearch) && (
)}
{/* Results count */}
Showing {unenrichedPage * UNENRICHED_PAGE_SIZE + 1}–{Math.min((unenrichedPage + 1) * UNENRICHED_PAGE_SIZE, filteredUnenriched.length)} of {filteredUnenriched.length} products
{unenrichedFilter !== 'all' && · filtered by {EAGLE_DATA.catalogEnrichment.unenrichedProducts.reasons.find(r => r.code === unenrichedFilter)?.label}}
{unenrichedSearch && · matching "{unenrichedSearch}"}
Page {unenrichedPage + 1} of {unenrichedTotalPages}
| Product | Brand | Category | GTIN | Existing Fields | Reason |
|---|---|---|---|---|---|
| No products match your current filters | |||||
| {p.name} | {p.brand} | {p.category} | {p.gtin} | = 30 ? 'text-emerald-400' : p.existingFields >= 10 ? 'text-yellow-400' : 'text-red-400'}`}> {p.existingFields} | {reasonMeta?.label || p.reason} |
{Array.from({ length: unenrichedTotalPages }, (_, i) => (
))}
)}
{/* Remediation callout */}
No GTIN Match
Request GTINs from store-brand team, or match by product name + brand against fuzzy lookup in USDA/Open Food Facts.
Insufficient Source Data
GTIN resolved but external sources returned incomplete or conflicting data below the confidence threshold. Re-process with additional data providers or manual review.
Discontinued
Archive or flag in catalog. Can be bulk-enriched if historical analytics (e.g., substitution modeling) require the data.
How Food Intelligence Powers the Agent
{[
{
title: 'Auto Health Badges',
desc: 'Agent auto-annotates search results with Food Compass, Nutri-Score, and NOVA badges so shoppers see health context instantly.',
icon:
{item.icon}
{item.title}
{item.desc}
70,000+
Total Products
68.3%
With Nutrition
+15.4% from enrichment
42.1%
Health Scores
+42.1% from Tufts
71.2%
With Ingredients
+12.7% from enrichment
Field Coverage Analysis
{[
{ field: 'Name', coverage: 100, color: 'bg-emerald-500' },
{ field: 'Category', coverage: 99.8, color: 'bg-emerald-500' },
{ field: 'Brand', coverage: 94.6, color: 'bg-emerald-500' },
{ field: 'Description', coverage: 87.3, color: 'bg-emerald-500' },
{ field: 'Ingredients', coverage: 71.2, color: 'bg-yellow-500' },
{ field: 'Nutrition Facts', coverage: 68.3, color: 'bg-yellow-500' },
{ field: 'Health Labels', coverage: 62.8, color: 'bg-yellow-500' },
{ field: 'Serving Size', coverage: 58.4, color: 'bg-yellow-500' },
{ field: 'Calories Data', coverage: 65.7, color: 'bg-yellow-500' },
{ field: 'Health Scores', coverage: 42.1, color: 'bg-orange-500' },
].map(item => (
))}
{item.field}
{item.coverage}%
{Math.round(70000 * (100 - item.coverage) / 100).toLocaleString()} gap
Enrichment Data Sources
{[
{ name: 'USDA FoodData Central', type: 'Government', coverage: '500K+ products', status: 'Primary', color: 'border-emerald-500', label: 'GOV' },
{ name: 'Open Food Facts', type: 'Open Source', coverage: '4M+ products', status: 'Active', color: 'border-blue-500', label: 'OSS' },
{ name: 'Tufts Food Compass', type: 'Academic', coverage: 'Scoring engine', status: 'Ready', color: 'border-purple-500', label: 'EDU' },
{ name: 'Edamam API', type: 'Commercial', coverage: '2.3M+ recipes', status: 'API Ready', color: 'border-yellow-500', label: 'API' },
{ name: 'Nutritionix', type: 'Commercial', coverage: '600K+ UPCs', status: 'API Ready', color: 'border-pink-500', label: 'API' },
{ name: 'Recipe1M+', type: 'Research', coverage: '1M+ recipes', status: 'Dataset', color: 'border-orange-500', label: 'R&D' },
].map(source => (
))}
{source.label}
{source.name}
{source.type}
{source.coverage}
{source.status}
Health Scoring Framework
Tufts Food Compass
Scores 1-100 based on 54 attributes across 9 health domains
70-100:Encourage
31-69:Moderate
1-30:Limit
Nutri-Score
European front-of-pack label from A (best) to E (worst)
{['A', 'B', 'C', 'D', 'E'].map((grade, i) => (
{grade}
))}
NOVA Classification
Food processing level from 1 (minimal) to 4 (ultra-processed)
Group 1:Unprocessed
Group 2:Culinary
Group 3:Processed
Group 4:Ultra-processed
{/* Context JSON Viewer */}
{/* Sample Conversation */}
{/* Context Signals Used */}
)}
{/* ================================================================= */}
{/* TAB — User Intelligence */}
{/* ================================================================= */}
{activeTab === 'userIntelligence' && (
Agent Context Structure
This is the actual context payload assembled for the LLM at the start of each conversation turn. Every tool call enriches this context further.
{`{
"shopper_profile": {
"shopper_id": "40001000512",
"mission_cluster": "health_focus",
"loyalty_tier": "MyPerks Gold"
},
"dietary_propensities": {
"organic": 0.91,
"gluten_free": 0.78,
"high_protein": 0.89,
"low_sodium": 0.82,
"high_fiber": 0.76,
"low_sugar": 0.71,
"dairy_free": 0.65,
"vegetarian": 0.34,
"nut_free": 0.15,
"vegan": 0.12,
"keto": 0.08
},
"virtual_pantry": [
{ "product_name": "Olive Oil", "stock_level": 0.65, "days_until_reorder": 12 },
{ "product_name": "Almond Milk", "stock_level": 0.72, "days_until_reorder": 15 },
{ "product_name": "Brown Rice", "stock_level": 0.45, "days_until_reorder": 6 },
{ "product_name": "Chicken Breast", "stock_level": 0.12, "days_until_reorder": 1 },
{ "product_name": "Greek Yogurt", "stock_level": 0.08, "days_until_reorder": 0 }
],
"active_promos": [
{ "promo_id": "P2024-1182", "type": "BOGO", "product": "Organic Pasta", "expires": "2024-12-15" },
{ "promo_id": "P2024-1195", "type": "20% OFF", "product": "Almond Milk", "expires": "2024-12-20" }
],
"session": {
"conversation_id": "conv_abc123",
"turn_count": 3,
"tools_called": ["search_products", "get_dietary_propensities"],
"cart_items": 2,
"cart_total": 18.47
}
}`}
Sample Conversation: Context-Driven Personalization
{/* User message */}
{/* Agent internal reasoning */}
{/* Agent response */}
U
I need ingredients for a pasta dinner
Agent Internal Reasoning
search_products
Query: "pasta dinner ingredients" — returns pasta, sauce, cheese, chicken, vegetables
get_dietary_propensities
gluten_free: 0.78
triggers dietary warning for regular pasta. Agent flags this and searches for gluten-free alternatives.
get_virtual_pantry
Olive Oil: 0.65 (in stock)
— excluded from shopping list since shopper already has it.
optimize_shopping_list
Removes pantry duplicates, applies active promos (20% off Almond Milk), sorts by aisle.
Here is your pasta dinner shopping list! I noticed a few things:
Based on your preferences, I've included gluten-free pasta instead of regular pasta. Your gluten-free preference score is high (0.78).
{[
{ name: 'Barilla GF Penne', price: '$3.49', note: 'Gluten-free' },
{ name: 'Rao\'s Marinara', price: '$7.99', note: 'NOVA Group 1' },
{ name: 'Organic Chicken Breast', price: '$8.99', note: 'High protein' },
{ name: 'Fresh Basil', price: '$2.49', note: 'Organic available' },
].map((item) => (
))}
{item.name}
{item.price}
{item.note}
I see you already have Olive Oil in stock (65%), so I've left it off the list. Your Greek Yogurt is running very low — want me to add it?
Context Signals Used in This Conversation
{[
{
signal: 'Dietary Propensities',
value: 'gluten_free: 0.78',
action: 'Swapped regular pasta for GF alternative, added warning badge',
color: 'border-yellow-500',
},
{
signal: 'Virtual Pantry',
value: 'Olive Oil: 0.65',
action: 'Excluded from list (already in stock); flagged Greek Yogurt for reorder',
color: 'border-emerald-500',
},
{
signal: 'Mission Cluster',
value: 'Health Focus',
action: 'Prioritized organic and clean-label options in search results',
color: 'border-purple-500',
},
{
signal: 'Active Promos',
value: '20% off Almond Milk',
action: 'Surfaced promo for pantry item running low on stock',
color: 'border-pink-500',
},
].map((s) => (
))}
{s.signal}
{s.value}
{s.action}
{/* ── Propensity Graph ──────────────────────────────────────────── */}
{/* Brand Affinities */}
{/* Behavioral Signals */}
{/* How Propensity Scores Are Computed */}
{/* ── Flavor DNA ────────────────────────────────────────────────── */}
{/* Computation Note */}
{/* Data Quality Impact Callout */}
)}
{/* ================================================================= */}
{/* TAB — Knowledge Graph */}
{/* ================================================================= */}
{activeTab === 'knowledgeGraph' && (
Shopper Propensity Model
11+ dimensions of preference scoring derived from 2-year purchase history joined against enriched catalog data
{/* Dietary Propensities */}Dietary Propensities
{EAGLE_DATA.userIntelligence.propensityDimensions.filter(d => d.category === 'dietary').map((d) => {
const barColor = d.score >= 0.7 ? '#10b981' : d.score >= 0.4 ? '#f59e0b' : '#64748b';
return (
);
})}
{d.name}
{d.score.toFixed(2)}
{d.score >= 0.7 && HIGH}
{d.score >= 0.4 && d.score < 0.7 && MED}
{d.score < 0.4 && LOW}
Brand Affinities
{EAGLE_DATA.userIntelligence.propensityDimensions.filter(d => d.category === 'brand').map((d) => {
const barColor = d.score >= 0.7 ? '#3b82f6' : d.score >= 0.4 ? '#60a5fa' : '#64748b';
return (
);
})}
{d.name}
{d.score.toFixed(2)}
{d.score >= 0.7 && HIGH}
{d.score >= 0.4 && d.score < 0.7 && MED}
{d.score < 0.4 && LOW}
Behavioral Signals
{EAGLE_DATA.userIntelligence.propensityDimensions.filter(d => d.category === 'behavioral').map((d) => {
const barColor = d.score >= 0.7 ? '#f59e0b' : d.score >= 0.4 ? '#fbbf24' : '#64748b';
return (
);
})}
{d.name}
{d.score.toFixed(2)}
{d.score >= 0.7 && HIGH}
{d.score >= 0.4 && d.score < 0.7 && MED}
{d.score < 0.4 && LOW}
How Propensity Scores Are Computed
{EAGLE_DATA.userIntelligence.propensityComputation.map((s, i) => (
{s.step}
{s.name}
{s.desc}
{i < 3 &&Flavor DNA Profile
Unique taste fingerprint derived from ingredient extraction across all purchased products
{/* Flavor Dimensions */}
{/* Cuisine Affinity */}
{/* Top Ingredients */}
Flavor Dimensions
{EAGLE_DATA.userIntelligence.flavorDNA.flavorDimensions.map((f) => (
))}
{f.name}
{(f.value * 100).toFixed(0)}%
Cuisine Affinity
{EAGLE_DATA.userIntelligence.flavorDNA.cuisineAffinity.map((c) => (
))}
{c.name}
{c.pct}%
Top 20 Ingredient Preferences
{EAGLE_DATA.userIntelligence.flavorDNA.topIngredients.map((ing, i) => (
{ing}
))}
Every product purchased is decomposed into its ingredient list via enriched catalog data. Ingredients are mapped to flavor compounds (sweet, salty, spicy, umami, sour, bitter) using our food ontology. Purchase frequency weights each ingredient, producing a per-shopper flavor fingerprint that powers cuisine recommendations and novel product discovery.
Predictive Pantry Model
ML-estimated stock levels, consumption rates, and reorder predictions for each shopper's household
{EAGLE_DATA.userIntelligence.pantryModel.items.map((item) => {
const urgencyColor = item.reorderIn <= 1 ? '#ef4444' : item.reorderIn <= 5 ? '#f59e0b' : '#10b981';
const urgencyLabel = item.reorderIn <= 1 ? 'Urgent' : item.reorderIn <= 5 ? 'Soon' : 'Stocked';
return (
);
})}
{item.name}
Rate: {item.consumptionRate}
Reorder in:
{item.reorderIn}d
{urgencyLabel}
conf: {(item.confidence * 100).toFixed(0)}%
{/* Household Size */}
{/* Seasonal Adjustments */}
Household Size Inference
{EAGLE_DATA.userIntelligence.pantryModel.householdSize.estimated}
estimated members
{EAGLE_DATA.userIntelligence.pantryModel.householdSize.method}
Confidence: {(EAGLE_DATA.userIntelligence.pantryModel.householdSize.confidence * 100).toFixed(0)}%
Seasonal Adjustments
{EAGLE_DATA.userIntelligence.pantryModel.seasonalAdjustments.map((s) => (
))}
{s.season}
{s.change}
Consumption Model
Inferred nutritional intake patterns from purchase history and household modeling
{/* Weekly Calories */}
{/* Meal Patterns */}
Estimated Weekly Caloric Intake
~{Math.round(EAGLE_DATA.userIntelligence.consumptionModel.weeklyCalories / 7).toLocaleString()} cal/day per household member
Macro Nutrient Balance
{EAGLE_DATA.userIntelligence.consumptionModel.macroBalance.map((m) => (
{m.pct}%
))}
{EAGLE_DATA.userIntelligence.consumptionModel.macroBalance.map((m) => (
{m.macro}
))}
{EAGLE_DATA.userIntelligence.consumptionModel.mealPatterns.map((m) => (
))}
{/* Health Trajectory */}
{m.meal}
{m.intensity}
{m.typicalCal}
avg cal · {m.pct}% of daily
{EAGLE_DATA.userIntelligence.consumptionModel.healthTrajectory.indicators.map((ind) => (
))}
{ind.change}
{ind.label}
Householding Model
De-averaging from loyalty card to individual household members using purchase pattern analysis and GNN inference
{/* Household Header */}Household ID
{EAGLE_DATA.userIntelligence.householding.householdId}
Confidence
{(EAGLE_DATA.userIntelligence.householding.confidence * 100).toFixed(0)}%
Members Detected
{EAGLE_DATA.userIntelligence.householding.members.length}
{Object.entries(EAGLE_DATA.userIntelligence.householding.detections).map(([key, val]) => (
))}
{key}
{val ? 'Yes' : 'No'}
{EAGLE_DATA.userIntelligence.householding.members.map((member) => (
))}
{/* Shared vs Personal Preferences */}
{member.role}
{member.identified ? 'Identified' : 'Inferred'}
Age: {member.age}
Preferences: {member.preferences}
ID: {member.shopperId}
Shared Preferences
{EAGLE_DATA.userIntelligence.householding.sharedPreferences.map((p) => (
{p}
))}
Primary Member Only
{EAGLE_DATA.userIntelligence.householding.personalPreferences.primary.map((p) => (
{p}
))}
Secondary Member Only
{EAGLE_DATA.userIntelligence.householding.personalPreferences.secondary.map((p) => (
{p}
))}
A single loyalty card often represents multiple people with different tastes. By de-averaging to the individual level, propensity scores become dramatically more accurate — the health-conscious primary shopper gets organic recommendations while the convenience-oriented secondary member gets quick-meal suggestions. This lifts personalization accuracy from ~60% (household-averaged) to ~94% (individual-level).
Data Quality & Anomalies
Of {EAGLE_DATA.userIntelligence.dataQuality.summary.totalLoyaltyUsers.toLocaleString()} loyalty users,{' '} {EAGLE_DATA.userIntelligence.dataQuality.summary.flaggedUsers.toLocaleString()}{' '} ({EAGLE_DATA.userIntelligence.dataQuality.summary.flaggedPct}%) are flagged with anomalies that could degrade model accuracy if left unresolved.
{/* Summary bar */}
{EAGLE_DATA.userIntelligence.dataQuality.anomalyTypes.map((a) => (
))}
{(100 - EAGLE_DATA.userIntelligence.dataQuality.summary.flaggedPct).toFixed(1)}% Clean
{EAGLE_DATA.userIntelligence.dataQuality.anomalyTypes.map((a) => (
))}
{/* Flagged Accounts Table */}
{a.label}
{a.severity}
{a.count.toLocaleString()}
{a.desc}
Impact
{a.impact}
Remediation
{a.remediation}
Flagged Account Examples
| Loyalty ID | Issue | Detail | mParticle IDs | Weekly Txns | Event Rate |
|---|---|---|---|---|---|
| {ex.loyaltyId} | {anomaly?.label || ex.issue} | {ex.detail} | 1 ? 'text-red-400 font-semibold' : 'text-gray-400'}>{ex.mparticleIds} | 10 ? 'text-amber-400 font-semibold' : 'text-gray-400'}>{ex.weeklyTxns} | {ex.eventRate} |
Without data quality hygiene, the {EAGLE_DATA.userIntelligence.dataQuality.summary.flaggedPct}% of flagged accounts can disproportionately skew model outputs. A single cashier card with 47 transactions/week pollutes propensity distributions more than 100 normal shoppers. Identity fragmentation (multiple mParticle IDs) dilutes individual taste signals, causing the flavor DNA and consumption models to under-represent preferences. Resolving these anomalies before model training is critical for production-grade personalization.
{/* ── Graph Architecture ────────────────────────────────────────── */}
)}
Knowledge Graph Architecture
A unified neural graph combining catalog, transaction, customer, behavioral, and enrichment data into a single queryable intelligence layer
{[
{ label: 'Total Nodes', value: EAGLE_DATA.knowledgeGraph.stats.totalNodes, color: 'from-blue-500 to-blue-600' },
{ label: 'Total Edges', value: EAGLE_DATA.knowledgeGraph.stats.totalEdges, color: 'from-emerald-500 to-emerald-600' },
{ label: 'Node Types', value: EAGLE_DATA.knowledgeGraph.stats.nodeTypeCount, color: 'from-purple-500 to-purple-600' },
{ label: 'Edge Types', value: EAGLE_DATA.knowledgeGraph.stats.edgeTypeCount, color: 'from-orange-500 to-orange-600' },
{ label: 'Avg Degree', value: EAGLE_DATA.knowledgeGraph.stats.avgDegree, color: 'from-pink-500 to-pink-600' },
{ label: 'Density', value: 'Sparse', color: 'from-cyan-500 to-cyan-600' },
].map((m) => (
))}
{/* ── Node & Edge Types ─────────────────────────────────────────── */}
{m.label}
{m.value}
{/* Node Types */}
{/* Edge Types */}
{/* ── Graph Traversal Examples ──────────────────────────────────── */}
Node Types
| Type | Count |
|---|---|
|
{n.type}
|
{n.displayCount} |
Edge Types
| Relationship | From → To | Count |
|---|---|---|
{e.type}
|
{e.from} → {e.to} | {e.count} |
Example Graph Traversals
{EAGLE_DATA.knowledgeGraph.queryExamples.map((q) => (
))}
{q.name}
{q.path}
{q.desc}
Value of the Knowledge Graph
The graph isn't just a data store — it's an intelligence multiplier that enables capabilities impossible with flat tables
{EAGLE_DATA.knowledgeGraph.graphValue.map((v) => (
))}
{v.title}
{v.desc}
Data Sources → Graph Integration
{EAGLE_DATA.knowledgeGraph.dataSources.map((ds) => (
))}
{ds.source}
{ds.records}
Creates: {ds.creates}
Edges: {ds.edgesCreated}